JP6932353B2 - Spray condition information calculation device, spray condition information calculation method - Google Patents

Description

The present invention relates to a spray condition information calculation device and a spray condition information calculation method.

Conventionally, the spraying method has been used for cleaning and painting. For example, a surface treatment is performed on a construction target such as a wall surface by a cleaning operation, and a coating film is sprayed. For example, Patent Document 1 describes the relationship between the physical properties of the paint sprayed on the vehicle and the amount of air supplied to the hand gun, which is to be painted on the vehicle, and is used for painting by referring to this. It controls the amount of air according to the paint.

Japanese Unexamined Patent Publication No. 2013-059757

However, in general, the construction method by spraying causes uneven construction depending on the construction technique and experience of the worker, the properties of the material to be sprayed, etc., and it is not always possible to perform the construction under the optimum conditions. There is a problem that it may be difficult to improve. In Patent Document 1, the amount of air supplied to the hand gun can be controlled according to the physical characteristics of the paint. However, since the construction by spraying cannot always be optimized only by the amount of air, there is room for improvement in Patent Document 1.

The present invention has been made in view of such circumstances, and an object of the present invention is to provide a spraying condition information calculation device and a spraying condition information calculation method capable of reducing unevenness in construction due to spraying.

In order to solve the above-mentioned problems, in the present invention, the cleaning process is performed in an environment according to the spraying state information including the spraying pressure, the spraying angle and the spraying flow rate of the liquid to be sprayed in the cleaning process, and the spraying state information. An acquisition unit that acquires adhesion performance information indicating the adhesion performance of the coating material to the object, which was constructed by spraying the coating material onto the target surface after being sprayed, and the spraying state information and the adhesion performance. Using the information, the learning unit that learns the spraying condition information representing the spraying pressure, the spraying angle, and the spraying flow rate of the liquid so that the adhesion performance of the coating material to the object reaches a predetermined reference value, and the learning It has an output unit that outputs a value indicating a spraying pressure, a spraying angle, and a spraying flow rate based on the obtained result.

Further, the present invention is a spraying condition information calculation method in a spraying condition information calculation device, which includes spraying state information including a spraying pressure, a spraying angle, and a spraying flow rate of a liquid to be sprayed by an acquisition unit in a cleaning process, and the spraying. Obtained the adhesion performance information indicating the adhesion performance of the coating material to the object, which was constructed by spraying the coating material on the target surface after the cleaning process was performed in the environment according to the state information. , The learning unit uses the spraying state information and the adhesion performance information to determine the spraying pressure, spraying angle, and spraying flow rate of the liquid so that the adhesion performance of the coating material to the object reaches a predetermined reference value. This is a spraying condition information calculation method in which the spraying condition information to be represented is learned, and the output unit outputs a value indicating a spraying pressure, a spraying angle, and a spraying flow rate based on the learning result.

As described above, according to the present invention, the liquid spraying condition information in the cleaning process is learned so that the adhesion performance of the coating material reaches a predetermined reference value, and the learning result is output. Preferred spraying conditions can be determined. As a result, by spraying the coating material according to the required spraying conditions, it is possible to carry out the enforcement so as to meet certain standards, reduce the unevenness of the construction due to spraying that does not meet the standards, and perform the construction. It becomes easier to maintain and improve quality.

It is a figure which shows the construction procedure of the resin coating film by embodiment of this invention. It is a schematic block diagram explaining the function of the spraying condition information arithmetic unit. It is a flowchart explaining operation of a spraying condition information calculation apparatus 10. It is a figure which shows an example of the adhesion test result. It is a figure which shows an example of a three-layer neural network.

Hereinafter, the spraying condition information calculation device according to the embodiment of the present invention will be described with reference to the drawings. Here, before explaining the spraying condition information calculation device, the spraying construction in the present embodiment will be described. In the present embodiment, the spraying work will be described as an example in which a resin coating film is applied to the surface to be constructed by using a spraying device.
FIG. 1 is a diagram showing a procedure for applying a resin coating film according to an embodiment of the present invention.
In the present embodiment, the method of applying the resin coating film is a method of constructing the concrete structure 1 by applying the polyurea-based resin coating film 3 after cleaning the concrete surface 2a which is the surface of the concrete 2. Be done.
The surface to be constructed is the concrete surface 2a. The method of applying the resin coating film 3 includes a cleaning step by a water jet cleaning operation and a resin spraying step.
The water jet cleaning work is carried out by a work (cleaning step) of scraping the concrete surface 2a by injecting high-pressure water using the water jet nozzle 4. The resin spraying step is carried out by spraying the resin coating film 3 onto the concrete surface scraped by the cleaning step using the spray nozzle 5.
In the resin spraying step, a spraying method is used in which polyurea resin is sprayed onto the surface of a concrete structure to a thickness of several millimeters. By carrying out the resin spraying method after performing this water jet cleaning work, the impact resistance and durability of the structure 1 such as concrete can be improved.

As the resin coating film 3, a synthetic resin composed of a polyurethane resin or a polyurea resin, which is a compound formed by a chemical reaction between isocyanate and a curing agent composed of at least one of a polyol and an amine, is adopted. can do.

Here, in the cleaning process, the adhesion performance of the resin coating film 3 to the concrete surface 2a differs depending on the state after the construction surface (concrete surface 2a) is scraped. Therefore, it is necessary to clean the resin coating film 3 so that the adhesion performance satisfies a certain standard. In this keren, the state of the kelen surface can change depending on various conditions such as the properties of the structure 1 and the concrete surface 2a and the method of spraying water sprayed from the water jet. Focusing on the spraying pressure, spraying angle, and spraying flow rate of the water jetted from the jet, it was decided to obtain these suitable conditions with the spraying control support device.

FIG. 2 is a schematic block diagram illustrating the function of the spray condition information calculation device.
The spray condition information calculation device 10 includes an input unit 11, a learning unit 12, and an output unit 13.
The input unit 11 is provided with spraying state information including the spraying pressure, the spraying angle, and the spraying flow rate of the liquid to be sprayed in the cleaning process, and the coating material is sprayed on the target surface after the cleaning process is performed according to the spraying conditions. Acquires the adhesion performance information indicating the adhesion performance of the applied coating material to the object. This liquid is, for example, water.

Further, the input unit 11 inputs data about the spraying state information including the spraying pressure, the spraying angle, and the spraying flow rate of the liquid to be jetted in the cleaning process via any input device such as a touch panel, a keyboard, or a mouse. Obtained by accepting input. This spraying state information will be described in the case where it is input via an input device. For example, a pressure sensor that measures the pressure of the liquid ejected by the water jet, the liquid ejected from the water jet, the injection target, and the angle. If data can be obtained from the flow sensor that measures the spray flow rate of the liquid jetted from the water jet as a result of calculating the spray angle, these data may be acquired.
The spraying angle, which is the angle between the liquid jetted from the water jet and the object to be sprayed, is determined by the angle of the water jet spraying direction with respect to the normal direction of the keren work surface. For example, by providing an attitude sensor on the water jet and measuring the angle of which direction the water jet is sprayed, the angle of the water jet with respect to the horizontal plane can be obtained, and the angle of the sprayed surface can be determined. , Obtained by measuring the angle with respect to the sprayed surface and the horizontal direction, or by reading from the construction drawing data. Then, by obtaining these relative angles from the angle of the injection direction of the water jet and the angle of the surface to be sprayed, the angle of the water jet spraying direction with respect to the normal direction of the keren work surface can be obtained. ..

The coating material is, for example, a synthetic resin used as a coating material for forming the above-mentioned resin coating film 3. Adhesion performance information is provided to the construction target surface of the resin coating film 3 after the resin coating film 3 formed by spraying the coating material on the construction target surface that has been chipped by the cleaning process is cured. Adhesion strength is evaluated and measured as adhesion performance. The input unit 11 acquires the adhesion performance information indicating the adhesion performance of the measured resin coating film 3.
The adhesion performance is measured, for example, by confirming the adhesion strength of the resin coating film when the resin coating film 3 is sprayed after the surface to be constructed is scratched by high-pressure water jetted by a water jet. This is done by evaluating the adhesion performance of the film. The input unit 11 accepts data input using this evaluation value as adhesion performance information.

The learning unit 12 uses the spraying state information and the adhesion performance information to represent the spraying pressure, the spraying angle, and the spraying flow rate of the liquid so that the adhesion performance of the coating material to the object reaches a predetermined reference value. Learn information. This learning unit is composed of, for example, a combination of a CPU (Central Processing Unit) of a computer or the like and a storage device (for example, a hard disk, RAM (Random Access Memory), ROM (Read Only Memory)), and the CPU becomes a storage device. The function as the learning unit 12 is realized by executing the program based on the stored learning algorithm.

The learning unit 12 neurally uses the adhesion result information representing the adhesion performance of the coating film formed on the surface cleaned by the cleaning process with respect to the spraying pressure, the spraying angle, and the spraying flow rate of the liquid sprayed in the cleaning process. Let the network learn.
In the learning unit 12, the learning model after the learning is performed by the learning is stored as a learned model. The trained model is used to output data showing the spray pressure, the spray angle, and the spray flow rate, which will be described later.

This learning is performed by reinforcement learning, but may be machine learning or the like.
As the learning model, for example, a neural network can be used. This neural network includes a first layer for calculating the spray pressure, a second layer for calculating the spray angle, and a third layer for calculating the spray flow rate.

Further, the learning unit 12 has a reward calculation unit 121 and a function update unit 122.
The reward calculation unit 121 calculates the reward based on the spraying state information input from the input unit 11 and the adhesion performance information input from the input unit 11.
The reward calculation unit 121 acquires the adhesion performance information input from the input unit 11 and calculates the reward for reinforcing learning the learning model in the learning unit 12.
The data used as a reward here is, for example, a value of adhesion performance. In reinforcement learning, for example, the learning device of the learning model can be updated so that the highest reward is obtained when the reference value of adhesion performance (1.5 N / mm ^{2) is exceeded.}
The function update unit 122 updates the function for calculating the spray condition information based on the spray state information acquired via the input unit 11 and the reward calculated by the reward calculation unit 121.

The output unit 13 outputs values indicating the spray pressure, the spray angle, and the spray flow rate based on the learning result. The output unit 13 is, for example, a liquid crystal display device, and may be output by displaying the value to be output on the screen, or may be output to an external device (for example, a spraying device) so as to be used as a set value. You may.

Such a spraying condition information calculation device 10 may be configured by, for example, a personal computer or a tablet, or may be provided in the spraying device.

Next, the operation of the spray condition information calculation device 10 described above will be described with reference to the flowchart of FIG.
The input unit 11 acquires the spraying state information by obtaining the information obtained from the input device or various sensors (step S101). The worker sprays water onto the wall surface or the like, which is the construction target surface, at the spraying pressure, spraying angle, and spraying flow rate of the water sprayed from the water jet according to the spraying state information. As a result, the cleaning process is carried out. When the cleaning process is carried out and the coating material is sprayed on the surface to be constructed, the adhesion performance is evaluated after a predetermined curing time has elapsed. When the adhesion performance is evaluated, the input unit 11 acquires the adhesion performance information representing the evaluation result (step S102). As the evaluation result, for example, a value representing the ^{adhesion strength (N / mm 2) is input as the adhesion performance.}

When the adhesion performance is acquired, the reward calculation unit 121 calculates the reward based on the spraying state information and the adhesion performance information (step S103).
For example, the reward calculation unit 121 calculates the reward so that the higher the adhesion performance, the higher the reward, and the lower the adhesion performance, the lower the reward. Further, the reward calculation unit 121 can also store a reference value representing the adhesion performance in advance and determine the reward based on the relationship between the reference value and the adhesion performance information. For example, if the adhesion strength represented by the adhesion performance information is less than the reference value, the reward will be increased as it approaches the reference value, and if the adhesion strength indicated by the adhesion performance information is greater than or equal to the reference value, the reward will be further increased. Can be calculated.
Next, the function update unit 122 updates the function for calculating the spray condition information based on the spray state information input from the input unit 11 and the reward calculated by the reward calculation unit 121 (step S104). ).

As described above, the spraying condition information calculation device 10 learns the spraying condition information.
When learning is performed, the output unit 13 of the spray condition information calculation device 10 outputs the spray condition information. This output may be displayed on the screen of the display device, for example, when the output unit 13 is a display device, or when the output unit 13 is an interface for outputting data to another device, spraying condition information. May be output as data. For example, when the output unit 13 is a display device, "spraying pressure XX MPa, spraying angle XX degrees, flow rate XX L / min" and the like are displayed on the screen as spraying condition information. The worker visually recognizes this and sets (or sets) the setting value of the water jet injection device so that the pressure and flow rate for injecting water from the water jet are set so as to be the spray state information corresponding to this spray condition information. Operate the valve, etc.) and maintain the injection angle so that it is the angle indicated by the spraying state information.
When the spraying condition information is output as data from the interface, for example, by outputting the spraying condition information to the spraying device that controls the pressure and flow rate of the water jetted from the water jet, the pressure is applied to the spraying device. The flow rate can be set.

Here, at the start of learning in the learning unit 12 (at the start of the learning algorithm), the correlation between the spraying condition information and the adhesion performance is unknown, but the learning is performed so that the adhesion performance satisfies a predetermined performance. Gradually identify the features and interpret the correlation as you proceed. Then, the learning unit 12 can bring the condition showing the correlation between the spraying condition information and the adhesion performance closer to the optimum solution as the learning progresses.

When performing a water jet cleaning operation on a certain construction target surface using such a spraying condition information calculation device 10, it is possible to obtain spraying condition information assuming that the following conditions are optimal.
Spray pressure: 150MPa ± 20MPa
Blow-out angle: 20 ° ± 5 ° from the normal of the Keren work surface
Spray flow rate: 6 L / min ± 0.5 L / min
Then, after the cleaning step is performed according to the spraying condition information, the adhesion performance when the resin coating film 3 is sprayed on the surface to be constructed is the adhesion performance when the resin coating film 3 is applied with a film thickness of 2 mm. It was confirmed that it exceeded the standard value (1.5 N / mm ^2).
The result of this adhesion test is shown in FIG. In this figure, the adhesion test results focusing on the relationship between the spray pressure of water and the adhesion performance among the spray condition information are shown. In this figure, the horizontal axis represents the water spraying pressure in the cleaning process, and the vertical axis represents the adhesion performance. Here, it was confirmed that ^{the adhesion performance (adhesion strength) exceeded the reference value (1.5 N / mm 2} ) when the spray pressure was between 100 MPa and 200 MPa.

As described above, the spraying conditions that satisfy the adhesion performance with respect to the required film thickness (for example, 2 mm) are based on the spraying pressure, the spraying angle, and the spraying flow rate, and these various amounts are set to three levels (for example, pressure). In this case, a three-layer neural network can be constructed as shown in FIG. 5 by dividing the pressure into a pressure larger than the average spray pressure, an average spray pressure, and a pressure smaller than the average spray pressure). By strengthening and learning the actual adhesion performance, bonus points are added to the level of each technical level, and by selecting and recording the optimum route, it becomes artificial intelligence by deep learning and optimal spraying construction can be performed. become able to.

In the embodiment described above, by using the above-mentioned spraying condition information calculation device 10, it is possible to standardize the spraying or optimize the spraying. For example, if the surface to be constructed has different properties, or if a coating material that has never been used is used, spraying condition information that meets the adhesion performance standard cannot be grasped. Even so, by advancing the learning by the spraying condition information calculation device 10, it is possible to obtain spraying condition information that satisfies the standard of adhesion performance.
Further, in the present embodiment, when artificial intelligence is used for the spraying condition information calculation device 10, the optimum spraying construction method is applied not only to the construction for forming a resin coating film but also to paint the body surface of a machine group or a vehicle. It can also be applied to a spraying factory or the like that forms a coating film.
Further, by forming the resin coating film 3 according to the spraying condition information that satisfies the adhesion strength, it is possible to reduce the waste of spraying the coating material more than necessary, and a water jet is required. Since it is possible to reduce the number of operations as described above, it is possible to reduce the electricity bill for operating the water jet.

Each part of the spraying condition information arithmetic unit 10 in the above-described embodiment may be realized by a computer. In that case, the program for realizing this function may be recorded on a computer-readable recording medium, and the program recorded on the recording medium may be read by the computer system and executed. The term "computer system" as used herein includes hardware such as an OS and peripheral devices. Further, the "computer-readable recording medium" refers to a portable medium such as a flexible disk, a magneto-optical disk, a ROM, or a CD-ROM, or a storage device such as a hard disk built in a computer system. Further, a "computer-readable recording medium" is a communication line for transmitting a program via a network such as the Internet or a communication line such as a telephone line, and dynamically holds the program for a short period of time. It may also include a program that holds a program for a certain period of time, such as a volatile memory inside a computer system that serves as a server or a client in that case. Further, the above program may be for realizing a part of the above-mentioned functions, and may be further realized for realizing the above-mentioned functions in combination with a program already recorded in the computer system. It may be realized by using a programmable logic device such as FPGA (Field Programmable Gate Array).

Although the embodiments of the present invention have been described in detail with reference to the drawings, the specific configuration is not limited to this embodiment, and includes designs and the like within a range that does not deviate from the gist of the present invention.

10 Spray condition information arithmetic unit 11 Input unit 12 Learning unit 13 Output unit 121 Reward calculation unit 122 Function update unit

Claims (3)

The coating material is sprayed on the target surface after the cleaning process is performed in the spraying state information including the spraying pressure, the spraying angle, and the spraying flow rate of the liquid to be sprayed in the cleaning process and the environment according to the spraying state information. The acquisition unit that acquires the adhesion performance information indicating the adhesion performance of the coating material to the object, which was constructed by being used,
Using the spraying state information and the adhesion performance information, spraying condition information indicating the spraying pressure, the spraying angle, and the spraying flow rate of the liquid so that the adhesion performance of the coating material to the object reaches a predetermined reference value can be obtained. Learning department to learn and
A spray condition information calculation device having a spray pressure, a spray angle, and an output unit that outputs a value indicating a spray flow rate based on the learned result. The spraying condition information calculation device according to claim 1, wherein the coating material is a resin. It is a spraying condition information calculation method in the spraying condition information calculation device.
For the spraying state information including the spraying pressure, the spraying angle, and the spraying flow rate of the liquid to be sprayed by the acquisition unit in the cleaning process, and the target surface after the cleaning process is performed in the environment according to the spraying state information. Acquires the adhesion performance information indicating the adhesion performance of the coating material to the object, which was constructed by spraying the coating material.
The learning unit uses the spraying state information and the adhesion performance information to represent the spraying pressure, the spraying angle, and the spraying flow rate of the liquid so that the adhesion performance of the coating material to the object reaches a predetermined reference value. Learn spraying condition information,
A spraying condition information calculation method in which an output unit outputs values indicating a spraying pressure, a spraying angle, and a spraying flow rate based on the learning result.

Images